Sleep-Wake Profile in Dementia with Lewy Bodies,Alzheimer’s Disease,and Normal Aging

Participants

Sixty-two patients affected by neurodegenerative dementia, 30 with a diagnosis of DLB (mean age: 77.52±4.89 years) and 32 diagnosed with probable AD (mean age: 76.92±8.31 years), were sequentially enrolled. Patients were recruited among those referred to the outpatient Memory Clinic at the Department of Neurosciences of the University of Padua (Italy). The diagnosis of DLB was made according to the consensus criteria recommended by the Consortium on DLB [13]. The diagnosis of probable AD was based on the current criteria of the National Institute on Aging-Alzheimer’s Association workgroups for the diagnosis of AD [15]. Exclusion criteria were: evidence of severe cerebrovascular disease on brain CT or MRI; history of psychiatric illnesses before dementia onset; Mini-Mental State Examination (MMSE) score <10/30 and absence of a reliable caregiver who could provide accurate information. Thirty-three elderly healthy participants (mean age: 73.19±7.64 years) recruited from the community and representative of brain aging without dementia served as control group. Controls were healthy elderlies with a MMSE score ≥27 recruited among community-dwelling volunteers. Controls were not excluded whether they reported some minor degree of sleep disturbances. Bed partners and caregivers of patients with dementia were excluded.

Clinical evaluation

Each patient underwent extensive general and neurological examination, and brain imaging (CT or MRI) in order to exclude severe cerebrovascular disease. Extrapyramidal signs were scored using the motor subsection (part III) of the Unified Parkinson’s Disease Rating Scale (UPDRS = score range: 0–128) [16]. The presence of fluctuations in cognition and alertness were assessed with the Clinician Assessment of Fluctuation (CAF) scale administered to the caregiver (score range: 0–12 and 16) [17]. Behavioral disturbances, such as delusions, hallucinations, agitation/aggression, dysphoria, anxiety, euphoria, apathy, disinhibition, irritability/lability, aberrant motor activity, sleep and eating disturbances, were investigated and graded with the Neuropsychiatric Inventory (NPI) questionnaire to the caregiver (score range: 0–144) [18]. The Charlson Comorbidity Index (CCI) [19] was used to evaluate the role of comorbidities. For this purpose, the condition of “dementia” was excluded from the CCI score. All participants underwent global cognitive assessment using the MMSE [20].

Sleep questionnaires and diaries

Daytime sleepiness was investigated using the Epworth Sleepiness Scale (ESS) validated in Italian language [21, 22]. The patient is asked to rate his/her likelihood of dozing off in 8 everyday situations using a scale ranging between 0 (never) and 4 (high probability to doze). A total score >11 is indicative of excessive daytime sleepiness.

The Pittsburgh Sleep Quality Index (PSQI) [23, 24], validated in Italian, was used to evaluate the night sleep quality. This is a 19-items questionnaire, which investigates subjective quality of sleep in the previous month. The first 4 questions investigate timing and length of sleep, while the following 15 questions require an answer that can vary between 0 and 3 (0 = the best), grading the occurrence of sleep problems, the entity of daytime dysfunction, the need for drugs to sleep, and the overall quality of sleep. The PSQI can distinguish between “good sleepers” and “poor sleepers”, where “poor sleepers” have a PSQI total score >5 (range 0–21). The PSQI total score derives from the sum of 7 sub-scores regarding specific areas that contribute to perceived sleep quality (duration of sleep, sleep disturbance, sleep latency, day dysfunction due to sleepiness, sleep efficiency, overall sleep quality, need meds to sleep).

Each participant was also asked about the experience of RBD using the REM Sleep Behavior Disorder Single-Question Screen (RBDQ1) [25]. Each of the previous questionnaires was explained and delivered to the patient, in presence of a caregiver, or to the bedpartner by an experienced neurologist (FF and GC).

Sleep diaries [26] were used to investigate usual sleep timing. Information was collected for 12 consecutive days on bed time, time try to sleep time, sleep latency, number of awakenings, wake up time, get up time, number of daytime naps, and their duration. The variables obtained from the diaries with a brief definition for each variable/derived variable are listed: 1) bed-time: time when the participant goes to bed; 2) try to sleep time: time when the participant actually attempts to commence sleep (as oppose to read, watch TV, etc., in bed); 3) sleep onset time: time when the participant actually falls asleep; sleep onset latency (SOL): time required to fall asleep (after attempting to commence sleep as opposed to after going to bed); 5) number of awakenings (NWAK); 6) wake up time: time when the participant wakes up; 7) get up time: time when the participant gets up; 8) wakefulness after sleep onset (WASO): time that the participant spends awake after sleep onset; 9) naps: number of daytime naps; 10) total sleep time (TST): the length of time between sleep onset time and wake up time; 11) time in bed (TIB): the length of time between bed time and get up time; 12) sleep efficiency (SE%): the proportion of sleep time respect to time in bed, obtained by calculating TST/TIB X 100.

Sleep diaries were completed only by patients who were considered well-enough to comply and/or whose caregiver could provide adequate daily supervision. In two instances (all single diaries within the set of DLB patients), it was not possible to quantify the number of daytime naps neither by the patients nor by the caregivers as daytime naps were too many and too pervasive during the day. These two cases were therefore excluded from data analysis.

Statistical analysis

The t test for independent samples, the Mann-Whitney U test and the χ² test (or Fisher’s exact test) were used for normal, ordinal, and categorical variables, respectively. Analysis of non-parametric variance was made using the Kruskall-Wallis test and the Mann Whitney U test as the post hoc test. Analysis of correlation was made using the Pearson’s r or Spearman’s ρ, as appropriate. The significance level was set at p < 0.05.

Clinical findings

Demographic and clinical characteristics of DLB, AD and control groups are shown in Table 1. DLB and AD patients were matched for age, gender, severity of global cognitive impairment, and severity of behavioral disorders. Healthy elderly participants were younger than AD (p = 0.04) and DLB (p = 0.002) patients and were more commonly females. The DLB group had significantly higher scores than AD in scales assessing parkinsonism and cognitive fluctuations, and experienced visual hallucinations more frequently. DLB had also the highest frequency of RBD-like symptoms (70% versus 9.4%). Moreover, DLB patients with RBD had a significantly higher CAF score compared to those without RBD (CAF score = RBD+: 5.21±3.70; RBD-: 2.11±3.22, p = 0.04) while no association was observed between the presence of RBD and visual hallucinations.

Patients with DLB and AD were on a higher number of psychoactive drugs compared to controls. Anti-dementia treatment (cholinesterase inhibitors and memantine) was equally distributed in DLB and AD groups. For behavioral disturbances, patients with AD were more commonly on antidepressants drugs and patients with DLB were more commonly on antipsychotic drugs. Hypnotic benzodiazepines were prescribed more commonly in controls (28%) compared to DLB (13%) and AD patients (6.3%). In DLB patients the most prescribed benzodiazepine was clonazepam, to treat RBD.

Findings from sleep questionnaires

Sleep quality and daytime sleepiness scores in the three populations are summarized in Table 2.

Daytime sleepiness (ESS) was significantly higher, almost twofold, in DLB compared to AD patients (p < 0.001) and controls (p < 0.001); ESS scores were abnormal in 23% of DLB patients compared to 3% of AD patients and 6% of controls. These data were not influenced by disease severity: correlation between ESS and MMSE was r = 0.09 for DLB (p = 0.6) and r = –0.27 for AD (p = 0.13).

Subjective night sleep quality (PSQI score) was significantly better (lower PSQI score) in AD patients compared to both DLB patients (p < 0.01) and controls (p < 0.01); PSQI scores were abnormal in 22.6% of AD patients versus 48.3% of DLB patients and 57.6% of controls. DLB patients had more pronounced daytime dysfunction due to sleepiness when compared to both AD patients (DLB: 1.06±1.11, AD: 0.41±0.76; p = 0.009) and controls (Controls: 0.48±0.71; p = 0.05). AD patients scored better than DLB patients and controls on several sub-items of the PSQI, including sleep efficiency (DLB: 1.03±1.27, AD: 0.26±0.63, Controls 0.84±1.11; AD versus DLB p = 0.009; AD versus Controls p = 0.03) and sleep quality (DLB: 0.93±0.88, AD: 0.47±0.57, Controls 1.04±0.68; AD versus DLB p = 0.02; AD versus Controls p = 0.001).

Once medication was taken into account, PSQI scores were not found to be associated with the use of hypnotics, antipsychotics, and anti-dementia drugs. ESS total scores in DLB were not substantially different between patients who were/were not on antipsychotics (p = 0.45), antidepressants (p = 0.05), non-benzodiazepine hypnotics (p = 0.19), benzodiazepines (p = 0.07), cholinesterase inhibitors (p = 0.3), and memantine (p = 0.9).

Findings from sleep diaries

Approximately 50% of patients affected by dementia provided adequately completed sleep diaries, compared to 100% of controls. Results for 16 DLB and 14 AD patients are shown in Table 3. In particular, time try to sleep and sleep onset time were significantly earlier in the DLB group compared to controls (DLB 22 : 49±00 : 54; Controls 23 : 33±00 : 40, p = 0.003 and DLB 23 : 10±00 : 52, Controls 23 : 48±00 : 42, p = 0.02, respectively). Since wake-up and get-up times were similar among groups, TST and TIB were significantly higher in the DLB group than controls (p = 0.01 and p = 0.02, respectively). The values of NWAK, WASO, SOL, and SE were comparable across groups. Two DLB patients experienced so many daily naps that could not be counted. We therefore excluded these two patients from further analysis. Although this constrains, the number of naps resulted significantly higher in DLB than in both AD (p = 0.03) and control (p = 0.03) groups. The sleep profiles of AD and controls were similar.

The TST did not correlate with CAF and ESS scores in the two disease groups (DLB: r = 0.3, p = 0.11; AD: r = 0.2, p = 0.27). As for the association with treatment, TST in DLB was associated with the use of clonazepam (mean hours: clonazepam+  = 10.2±0.8; clonazepam- = 7.8±1.2; p = 0.01) but not with other types of drugs (ChEI: p = 0.1; memantine: p = 0.7; antidepressants: p = 0.32; antipsychotics: p = 0.42).

Presence of RBD was not found significantly associated with more naps, although the mean value of number of naps was higher in RBD+ respect to RBD- (RBD+: 0.85 ± 0.87; RBD-: 0.45 ± 0.53; p = 0.12).

Sleep-wake profile in DLB

DLB patients are characterized by two distinct features that differentiate them from AD and controls. Firstly, a more pronounced impairment of wake, consisting of increased daytime sleepiness, frequent naps, and perceived daytime dysfunction due to sleepiness. Secondly, sleep was often disturbed with RBD, with a longer TST. The prevalence of RBD in the DLB group (70%) is in agreement with the reported estimated prevalence [7]. Differences between studies could be due to different assessment tools and heterogeneity of patients sample (i.e., mixed DLB and PD-dementia) [10, 11]. In our DLB group, the TST was associated with the use of clonazepam to treat parasomnia.

It is not completely understood whether inefficient sleep in DLB, due to intrinsic alteration of sleep structure [10, 11] and parasomnia, may influence daytime alertness or whether they are independent, specific features of DLB [27]. In this study, presence of RBD has been found to be associated with the severity of fluctuations in attention but not with day-time somnolence assessed with the ESS. This questionnaire may be not adequate to assess hypersomnolence in patients with moderate dementia while number of naps could better intercept this feature. Presence of RBD was associated with more naps, although the difference did not reach significance, possibly due to the small patients’ sample.

We found an association between presence of RBD and severity of fluctuation in attention. A previous study assessing clinical-pathological association of RBD in DLB found an association between presence of RBD and early appearance of visual hallucinations and parkinsonism, but not with presence of fluctuations [28]. Differently from this study, we considered severity of fluctuations instead of presence of fluctuations, which was evaluated with a different scale (CAF instead of Mayo Fluctuation Scale). The possible association of RBD with severity of fluctuations may suggest that DLB-related neuropathological changes may target anatomical structures involved in both RBD and fluctuations such as brainstem nuclei and its projections to the thalamus.

In synthesis, this picture could delineate a scenario in which neither sleep or wake is efficient enough to allow for an adequate global functioning of these patients. Further studies are required to investigate whether improving parasomnia and/or impaired wakefulness could have an impact on global cognitive performance of DLB patients, especially in the early stages of disease.

Sleep-wake profile in AD

The discrepancy between subjective evaluations of sleep quality and more objective data collected with daily sleep diaries in AD patients parallels the findings of a previous study [29] in which subjective reports were compared to data collected using actigraphy. The authors investigated sleep quality in a group of AD patients in the early-moderate stage and in age-matched control group and showed that reported sleep disturbance were underestimated when compared to actigraphic parameters. Underestimation of sleep disturbances in AD patients may represent an important clinical issue for physicians, since it could lead to under-treatment, with consequences on cognitive performance and mood, as suggested by Most et al. [29] and several other studies [30–33]. The hypothesis of under-treatment of sleep disturbances is somewhat supported by the observation, in our study, that AD patients showed the lowest rates of consumption of hypnotic medications. Another study reported a poor sleep quality in AD as well as other types of dementia such as vascular dementia and DLB [34]; however in this study, AD patients were more severely impaired with respect to our samples. An increased frequency of sleep disturbances in more severely compromised patients has been already reported [35].

The neuropsychological substrate underlying the underestimation of sleep disturbances in AD patients is not fully understood. Most and colleagues [29] suggested that impaired comprehension of written questionnaires, impaired recall of recent events, and poor ability to self-report internal events could all contribute. In our study, questionnaires were administered and explained by an examiner to patients with the assistance of the caregiver, and the answers recorded by the examiner him/herself. Thus, it is unlikely that impairment in written comprehension may have played a major role. The impact of impaired recall is also unlikely to play a major role because homogenous answers toward a good sleep quality were obtained in the whole AD population, while impaired recall would have probably lead to erratic, random answers (probably an equal proportion of good, bad, neutral). Deficiency in self-assessment and monitoring of internal events and anosognosia for disturbances related to the illness may have a crucial role. In this context, the day-by-day recall of sleep quality, as it is requested with the daily diaries, avoids the possible errors introduced by a single point question requiring a judgment on a general estimate of sleep quality.

Sleep-wake profile in healthy elderly controls

Healthy elderly people had negative opinions of their sleep quality and showed the shortest TST compared to both DLB and AD. Our data shows that in this population there are subjectively increased NWAK and SOL and decreased TST, suggestive of a burden of symptoms related to insomnia. It has been already demonstrated that PSQI scores change with age [36, 37]. Impaired perception of sleep quality in old age could be due to the comparison between the current sleep situation that underwent physiological impairment and the previous features of sleep in the younger age. This situation, along with the extension of life expectancy and, above all, the extension of the active period of life, could lead to disproportionate expectations of sleep-related restoration in the older age. The practical counterpart of this observation is the high request of drugs to improve sleep, as demonstrated by a significant proportion of healthy participants been prescribed sleep inducing benzodiazepines in our study. Finally, it should be noticed that the Control group recruited for this study is representative of community-dwelling healthy elderly individuals, and that minor sleep troubles were not an exclusion criteria. Therefore it does not represent a “pure” control condition.

Conclusion

Patients with DLB have severe daytime somnolence with frequents naps and night-time RBD. The relationship between day-time and night-time alterations is not yet elucidated and needs further studies, although an association between presence of RBD and severity of fluctuations in attention could suggest a more complex reciprocal influence. In contrast, self-reported sleep disturbances in AD are underestimated and could be therefore unrecognized and under-treated.

Limitations of this study are the relative small sample size of patients with DLB and AD and the lack of neuropathological confirmation of the diagnosis (although the last point was minimized by confirming the diagnosis after a follow-up period). Moreover, we described RBD-like symptoms and could not make a diagnosis of RBD since it requires polysomnography, an instrumental tool not easily available for elderly people with dementia. Finally, our group of elderly controls was younger than patients groups and composed of more females participants. These features could have influenced some sleep/wake variables or the prevalence of RBD in healthy participants.

Future research directions should address whether RBD or also other parasomnias are more frequent in DLB respect to AD; whether sleep disturbances in DLB co-occur with daytime hypersomnolence or influence each other in a vicious loop; whether improving wake results in a better sleep and vice-versa; which simple tools could better intercept these wake-sleep alterations in a population affected by dementia.

We suggest that an easy and inexpensive first level screen of sleep disturbances in people with dementia could be carried out with a combination of sleep questionnaires and diaries, as these clinical tools demonstrated to be complimentary in the detection of sleep disturbances in this population, due to the possible occurrence of underreported sleep troubles.